One Source Hash Can Have Several Consumers
Why a set of artifact identities is not a count of rewritten files or module instances.
Scope: Engineering or methodology note based on a September 17, 2026 snapshot. Source-level behavior is not a public availability or universal compatibility claim. Any scenarios are illustrative, not customer outcomes.
Content hashes are excellent identifiers. They can also be overinterpreted. If two consumer files contain identical captured bytes, a set of source-body hashes may contain one entry even though the operation affects more than one consumer.
KodeBack’s externalized parity records distinguish source artifact identities from consumer accounting. That distinction matters when reviewing a multi-consumer package replacement.
Deduplication does not remove obligations
Imagine a synthetic capture where two paths serve identical JavaScript bodies. Deduplicating their content identity is reasonable. It should not cause the planner or verifier to forget that both consumer paths need their own valid rewrite and binding.
The current source verifies each exact consumer path, captured body, locator cohort, rederived carve, and certificate binding before digest deduplication. The resulting source-artifact set describes unique body identities, not how many consumers were handled.
Candidate identity must describe the whole generation
For multiple consumers, candidate-generation identity is derived from a canonical collection of relative paths, byte lengths, and content hashes. The complete publication generation also binds the full served tree. This prevents a multi-file operation from being summarized by one convenient output hash that omits a changed consumer.
The practical value is not the hash format itself. It is the ability to ask whether every intended output belongs to the same reviewed generation and whether a later file change invalidates that claim.
Do not infer runtime identity from byte identity
Identical source bytes do not automatically prove one browser module instance. Different consumers, documents, or resolution contexts can still matter. The source-artifact set is not evidence of singleton behavior by itself.
That is why package-family and browser-resolution questions remain separate from content deduplication. A clean set of hashes can authenticate bytes without answering every execution relationship those bytes participate in.
This article describes the September source’s reporting and identity boundary. It does not claim that all multi-consumer or module-resolution cases are supported.
When reading a recovery report, ask what each count counts. Unique bodies, consumer paths, rewritten files, and runtime instances are different denominators. A trustworthy report should not let one number impersonate the others simply because hashes make the summary look exact.